A Similarity Analysis of DGGE Images using DP matching

In this paper we propose a new evaluation method of similarity between DNA samples in an image obtained by polymorphism analysis called DGGE (denaturing gradient gel electrophoresis) (Borresem et al., 1988 and Muuzer et al., 1993). DGGE uses an electrophoresis unit and needs to observe reactions of samples included in the gel. Therefore the samples are often visually distorted. Conventionally the analysis of the DGGE images has been performed as visual evaluations, which requires a lot of human labor. Moreover the reliability of the analysis highly depends on skills of the operator and the reproducibility the burden of human labor. The proposed method automates the process of extraction of data corresponding to samples from a DGGE image, and undergoes pattern matching with a degree of similarity for each sample. A distortion of DGGE is compensated by use of a DP matching, often used in the signal processing. Experimental results shows that A degree of similarity and classification between samples are enhanced effectively.